Specificity in Research on Human Lactation: Opportunities for Growth

We have spent 2024’s year of editorials defining breastfeeding research. This editorial will present some areas for growth and improvement. The goal of healthcare research, and in our case, research about every aspect of the human phenomenon of breastfeeding, is to saturate the evidence base for a particular topic. This provides an empirical body of knowledge that can continually be tested and updated to safely inform practice. For example, if we consider the concept of providing additional support and education to breastfeeding parents, the peer-reviewed/empirical literature has shown that if the education is adequate, the exclusivity, duration, and breastfeeding self-efficacy will be improved. We have approached education and support interventions from any number of variations in timing, types of providers giving the support and education, group or individual education modality, inclusion of partners or families, and use of innovative educational tools. While the acquisition of information has circular elements—meaning that there may be reason to continually test what we think we know—at this point, we can safely say that when parents are provided with more breastfeeding support and education, at the population level, we generally see behavioral changes that improve breastfeeding success.

This does not mean that there is nothing left to research in this topic area, but it does mean that the level of detail and specificity needed to contribute to the literature becomes more critical. JHL often receives papers that cover a well-researched topic area, such as assessing the effect of additional breastfeeding support versus standard hospital or clinical care on breastfeeding outcomes. Specificity, in this case, could take many forms. For example, the research might test an established, standardized intervention in a unique population, providing an in-depth discussion of the cultural and contextual outcomes. If, instead, the researcher develops an intervention unique to their individual setting rather than applying an existing standardized intervention, more is required. In these situations, complete and detailed descriptions of both the intervention and the control would help to make the study accessible to readers and comparable to other studies in the same general topic area. These papers would be further enhanced by a background section that clearly provides this information to readers rather than reiterating the generally agreed-upon need for further support of breastfeeding. Sometimes there is a particular tool that is being tested, such as the use of wearable models, an online application, or a unique format (e.g., groups or storytelling techniques). These are viable and interesting studies that can increase our knowledge of what specific elements of breastfeeding education are helpful. However, if the control group in these studies receives only standard care, breastfeeding support in general becomes difficult to disentangle from the specific tool being applied. If the tool is the innovation, the control arm should also receive breastfeeding education so that we isolate the effects of the tool specifically.

Another consideration in this type of study is experimenter bias, which is the bias that is introduced when the researcher’s expectations or attitudes unconsciously influence the participants or the outcomes of the study. If the intervention is both applied and assessed by the same researcher, particularly if that researcher is also a care provider, it becomes difficult to separate the influence of the relationship between breastfeeding parent and provider from the intervention itself. This is particularly important in lactation research because the process of breastfeeding, and particularly early breastfeeding, is a learned and socially influenced behavior. A specific area of concern is the follow-up assessment of breastfeeding. This process itself can elicit questions, the provision of support, or the uncovering of breastfeeding issues that need attention. If the assessment is carried out by a care provider who could provide that support but does not, that could be particularly disheartening for the participant.On the other hand, if support is provided during the assessment, that needs to be accounted for when evaluating the outcomes. This can be considered similar to the Hawthorne Effect, where the modification of behavior is altered in response to being observed. However, in the context of breastfeeding research in which care providers are also the researchers, the effect of the relationship and the expectations of participants may be larger than what would be observed in a study in which a relationship has not been previously established between researcher and participant.

Part of the reason these considerations are important is scientific; that is, careful consideration of the study design, the research question, and how to provide the specificity and accessibility to study criteria needed to add to the evidence base. The other part is due to the nature of breastfeeding. It is human behavior, and a complex one, that encompasses the vulnerability of parenthood, the social context in which it occurs, and the sense of responsibility we feel for the survival of our children. All of these can introduce bias to study outcomes and are worth considering when planning study interventions.

Breastfeeding exclusivity, duration, and self-efficacy provide valuable broad measures of breastfeeding success, but many datasets allow us to refine our observations by creating more specificity in the outcome variables (e.g., the volume of milk produced, happiness with breastfeeding, early and unintentional weaning) or by designing studies that drill down to unique and very specific areas of lactation (e.g., the difference between an intervention offered during the hospital stay or the same intervention offered at one week postpartum). The challenge in presenting explicit, fine-tuned outcomes is to stay true to the study design without being led by the statistical strength of the results during the study analysis. The outcomes defined during the construction of the research question (a priori) should stay the same throughout the methodological approach and the discussion section. For example, if simple regression (one independent and one dependent variable) is performed to assess the strength of associations between the independent variables and the outcome without further modeling to account for confounding factors (variables associated with both the exposure and the outcome), the statistically significant associations will often be factors well-established in the literature—the things we already know to affect breastfeeding outcomes. It is easy, particularly for new researchers, to be sidetracked by the strength of associations at this stage of the data analysis, and thus be distracted from the potential unique or novel outcomes that could be explored with their data. One example might be a study that is well set up to assess the use of a wearable breastfeeding model during breastfeeding education, but the authors focus on the importance of breastfeeding intention as the strongest predictor of their participants’ breastfeeding success outcome. In this situation, they have diminished the ability of the research to contribute to the literature by focusing on the strongest statistical predictors instead of building a model that controls for the influence of intention to breastfeed, allowing them to remain true to the study’s original purpose. The most robust studies will have a very clear storyline, from the research question to the results and through the discussion and conclusion. The most engaging results will either be carefully constructed and well-defined so that they are comparable to other breastfeeding research, or will artfully dissect breastfeeding success to offer distinct findings that expand our understanding of the intricacies of what breastfeeding is and feels like to the participant, or how it is supported by family, care providers, institutions, policies, or cultural expectations.

There are still many gaps in the breastfeeding literature, and our authors work hard to uncover and fill those gaps. The heart of our role as a journal in human lactation is to continue to expand and extend our knowledge base. As editors, doing so requires us to pay particular attention to study design and outcomes to ensure complete and explicit communication between the researcher/author and the reader of the study manuscript. The system of peer review and editorial support is built on the precept that, with the support of the lactation research community, we can, together, continue to build our understanding of all the elements of breastfeeding as a biological and social phenomenon.